Shift Imaging Research Articles

Variational Color Shift and Auto-Encoder Based on Large Separable Kernel Attention for Enhanced Text CAPTCHA Vulnerability Assessment

Text CAPTCHAs are crucial security measures deployed on global websites to deter unauthorized intrusions. The presence of anti-attack features incorporated into text CAPTCHAs limits the effectiveness of evaluating them, despite CAPTCHA recognition being an effective method for assessing their security. This study introduces a novel color augmentation technique called Variational Color Shift (VCS) to boost the recognition accuracy of different networks. VCS generates a color shift of every input image and then resamples the image within that range to generate a new image, thus expanding the number of samples of the original dataset to improve training effectiveness. In contrast to Random Color Shift (RCS), which treats the color offsets as hyperparameters, VCS estimates color shifts by reparametrizing the points sampled from the uniform distribution using predicted offsets according to every image, which makes the color shifts learnable. To better balance the computation and performance, we also propose two variants of VCS: Sim-VCS and Dilated-VCS. In addition, to solve the overfitting problem caused by disturbances in text CAPTCHAs, we propose an Auto-Encoder (AE) based on Large Separable Kernel Attention (AE-LSKA) to replace the convolutional module with large kernels in the text CAPTCHA recognizer. This new module employs an AE to compress the interference while expanding the receptive field using Large Separable Kernel Attention (LSKA), reducing the impact of local interference on the model training and improving the overall perception of characters. The experimental results show that the recognition accuracy of the model after integrating the AE-LSKA module is improved by at least 15 percentage points on both M-CAPTCHA and P-CAPTCHA datasets. In addition, experimental results demonstrate that color augmentation using VCS is more effective in enhancing recognition, which has higher accuracy compared to RCS and PCA Color Shift (PCA-CS).

Numerical simulation and experimental study on guidance performance of hypersonic seeker under aerodynamic optical transmission effects

When a hypersonic seeker flies at high speed within the atmosphere, intense interaction with the incoming flow gradually develops into a complex turbulent flow field. This interaction results in complex thermal responses at the seeker window, causing aerodynamic optical effects such as image shift, jitter, and blur of the target image, thereby restricting the seeker's detection capability and accuracy. This paper uses a numerical simulation model for the guidance performance of a hypersonic seeker under aerodynamic optical transmission effects. The study focuses on an ellipsoidal seeker, with its supersonic flight simulation on the basis of the Reynolds-Averaged Navier-Stokes (RANS) equations to get a non-uniform gradient flow field. The correctness of the flow filed results can be verified by wind tunnel experiments. The transient temperature field of the seeker is solved using an unsteady thermal conduction-radiation coupled fluid-solid heat transfer method. Finally, the guidance performance of the hypersonic seeker under aerodynamic optical effects is predicted using the ray tracing method, which employs wavefront aberration, point spread function, degraded images, and image shift.

Mix-DDPM: Enhancing Diffusion Models through Fitting Mixture Noise with Global Stochastic Offset

Denoising diffusion probabilistic models (DDPM) have shown impressive performance in various domains as a class of deep generative models. In this article, we introduce the mixture noise-based DDPM (Mix-DDPM), which considers the Markov diffusion posterior as a Gaussian mixture model. Specifically, Mix-DDPM randomly selects a Gaussian component and then adds the chosen Gaussian noise, which can be demonstrated as a more efficient way to perturb the signals into a simple known distribution. We further define the reverse probabilistic model as a parameterized Gaussian mixture kernel. Due to the intractability in calculating the KL divergence between Gaussian mixture models, we derive a variational bound to maximize the likelihood, offering a concise formulation for optimizing the denoising model and valuable insights for designing the sampling strategies. Our theoretical derivation highlights that Mix-DDPM need only shift image which requires the inclusion of a global stochastic offset in both the diffusion and reverse processes , which can be efficiently implemented with just several lines of code. The global stochastic offset effectively fits a Gaussian mixture distribution enhancing the degrees of freedom of the entire diffusion model. Furthermore, we present three streamlined sampling strategies that interface with diverse fast dedicated solvers for diffusion ordinary differential equations, boosting the efficacy of image representation in the sampling phase and alleviating the issue of slow generation speed, thereby enhancing both efficiency and accuracy. Extensive experiments on benchmark datasets demonstrate the effectiveness of Mix-DDPM and its superiority over the original DDPM.

Putting the Mockers On: The Rutles, The Beatles, Rock Biopics and Parody

Many accounts of film genre reiterate a familiar narrative of growth to “classic” maturity and subsequent parody and/or deconstruction, and the biopic is no exception. However, rock music biopics have reversed this narrative, so that the genre begins in parody and only gets serious later. This is partly because rock and roll music began as parody, mainly by white people imitating African Americans, what is known as blackface minstrelsy, in which music and humour are necessarily (because of racism) mixed. In turn, the 60s rock counterculture took many of its cues from this untimely birth, appropriating African-American marginality in modes that were at once serious (concerns about authenticity) and ironic (mockery of Establishment values). This collision of opposites helps explain both the counterculture’s preference for documentary, especially of live performance, over Hollywood fiction, and its predilection for mockery of both (for example, mockumentary). As the single most influential proto-rock act, whose inventive wit and comic antics, rendered in newsreel, direct cinema, cartoon, and on record, were keys to their commercial and critical success, The Beatles were the perfect subjects for such ironic canonisation. Their filmic career highlights the intersection of documentary and comedy, as well as reality and fiction, via musical performance, a mode which can problematise documentary/comedy, and reality/fiction distinctions. In line with this argument, I have focused on key live performances from the Beatles’ career, and how they are parodied in The Rutles: All You Need Is Cash (1978) henceforth The Rutles, which doubles as the first filmic biography of the Beatles and the first rock mockumentary. The Beatles’ later career saw their public image shift from intentional to unintentional comedy, a shift mapped in the The Rutles, which gradually moves from parody towards satire. It is argued that The Rutles is open to a range of audience identifications and readings: it is at once a text for “true fans”, a playful deconstruction of their investments, but also one with real-world reverberations (some of its predictions came true). In this sense, it is a “media savvy”, peculiarly contemporary text that questions the priority of reality over fiction.

Magnetic Resonance Imaging-Based Monitoring of the Accumulation of Polyethylene Terephthalate Nanoplastics.

Polyethylene terephthalate (PET) is one of the most produced plastic materials in the world. The emergence of microplastics and nanoplastics (MPs/NPs) as a significant environmental contaminant has become a matter of increasing concern. While the toxicological effects of PET NPs have been widely researched, there is a lack of methodologies for studying their accumulation. The present study introduces a novel method to monitor the distribution of PET NPs in germinating wheat (Triticum aestivum L.) seeds. This involves the functionalization of superparamagnetic iron oxide nanoparticles (SPIONs) with PET NPs (PET-fSPIONs) coupled with magnetic resonance microimaging (µMRI) to provide insight into their distribution within the seed. The present study has demonstrated that PET-fSPIONs accumulate in specific regions of germinating wheat seeds, including the shoot apical meristem, the radicle, the coleoptile, the plumule, and the scutellum. Furthermore, the accumulation of PET-fSPIONs has been shown to exert a discernible effect on spin-spin relaxation (T2), as observed via MRI and quantitative T2 relaxation time analysis. The accumulation of PET NPs in embryo regions was also confirmed by SEM. Diffusion-weighted magnetic resonance imaging (DW-MRI) and non-invasive chemical shift imaging analyses demonstrated that PET NPs resulted in restricted diffusion within the highlighted areas, as well as an impact on lipid content. Our study reveals that using µMRI with fSPIONs provides a non-invasive method to monitor the biodistribution of PET nanoparticles in wheat seeds. Additionally, it offers valuable insights into the microstructural interactions of PET.

Aberration compensation in Doppler holography of the human eye fundus by subaperture signal correlation.

The process of obtaining images of capillary vessels in the human eye's fundus using Doppler holography encounters difficulties due to ocular aberrations. To enhance the accuracy of these images, it is advantageous to apply an adaptive aberration correction technique. This study focuses on numerical Shack-Hartmann, which employs sub-pupil correlation as the wavefront sensing method. Application of this technique to Doppler holography encounters unique challenges due to the holographic detection properties. A detailed comparative analysis of the regularization technique against direct gradient integration in the estimation of aberrations is made. Two different reference images for the measurement of image shifts across subapertures are considered. The comparison reveals that direct gradient integration exhibits greater effectiveness in correcting asymmetrical aberrations.

Gini Coefficient-Based Feature Learning for Unsupervised Cross-Domain Classification with Compact Polarimetric SAR Data

Remote sensing image classification usually needs many labeled samples so that the target nature can be fully described. For synthetic aperture radar (SAR) images, variations of the target scattering always happen to some extent due to the imaging geometry, weather conditions, and system parameters. Therefore, labeled samples in one image could not be suitable to represent the same target in other images. The domain distribution shift of different images reduces the reusability of the labeled samples. Thus, exploring cross-domain interpretation methods is of great potential for SAR images to improve the reuse rate of existing labels from historical images. In this study, an unsupervised cross-domain classification method is proposed that utilizes the Gini coefficient to rank the robust and stable polarimetric features in both the source and target domains (GRFST) such that an unsupervised domain adaptation (UDA) can be achieved. This method selects the optimal features from both the source and target domains to alleviate the domain distribution shift. Both fully polarimetric (FP) and compact polarimetric (CP) SAR features are explored for crop-domain terrain type classification. Specifically, the CP mode refers to the hybrid dual-pol mode with an arbitrary transmitting ellipse wave. This is the first attempt in the open literature to investigate the representing abilities of different CP modes for cross-domain terrain classification. Experiments are conducted from four aspects to demonstrate the performance of CP modes for cross-data, cross-scene, and cross-crop type classification. Results show that the GRFST-UDA method yields a classification accuracy of 2% to 12% higher than the traditional UDA methods. The degree of scene similarity has a certain impact on the accuracy of cross-domain crop classification. It was also found that when both the FP and circular CP SAR data are used, stable, promising results can be achieved.

The memories we (dis)like: wartime global news and Ukrainian-language poetry as semantic weapons

The research notes examine a shared memory created during the Russian war against Ukraine, as depicted in news coverage of global media outlets, and in Ukrainian-language competitive poetry from international free competitions and Ukrainian governmental art projects. The research goal is to explain the hidden dangers and consequences of war memorization. The author compares public collective evaluations of key metaphors used in news during the initial three weeks of war and analyzes key images in Ukrainian poetry from pre-war and wartime periods. Preliminary results of the secondary qualitative content analysis of news and analysis of group discussions with students specializing in different fields illustrate how the metaphors varied in terms of adequacy and artistic expressiveness due to societal evaluations. The findings of quantitative semantic analysis of Ukrainian-language poems highlight how the shift in key images correlates with the transition from a relatively peaceful context to an overtly military one. The knowledge gained contributes to a better understanding of memory weaponization and specific ways in which collective sensitivity is little influenced by media positions (pro-Ukrainian or non-pro-Ukrainian), being more dependent on a sense of morality.

B1-MRF: Large dynamic range MRF-based absolute mapping in the human body at 7T.

This study aims to map the transmit magnetic field ( ) in the human body at 7T using MR fingerprinting (MRF), with a focus on achieving high accuracy and precision across a large dynamic range, particularly at low flip angles (FAs). A FLASH-based MRF sequence (B1-MRF) with high sensitivity was developed. Phantom and in vivo abdominal imaging were performed at 7T, and the results were compared with established reference methods, including a slow but precise preparation-based method (PEX), saturated TurboFLASH (satTFL), actual flip angle imaging (AFI) and Bloch-Siegert shift (BSS). The MRF signal curve was highly sensitive to , while T1 sensitivity was comparatively low. The phantom experiment showed good agreement of to PEX for a T1 range of 204-1691 ms evaluated at FAs from 0° to 70°. Compared to the references, a dynamic range increase larger than a factor of two was determined experimentally. In vivo liver scans showed a strong correlation between B1-MRF, satTFL, and RPE-AFI in a low body mass index (BMI) subject (18.1 kg/m2). However, in larger BMI subjects (≥25.5 kg/m2), inconsistencies were observed in low regions for satTFL and RPE-AFI, while B1-MRF still provided consistent results in these regions. B1-MRF provides accurate and precise maps over a wide range of FAs, surpassing the capabilities of existing methods in the FA range < 60°. Its enhanced sensitivity at low FAs is advantageous for various applications requiring precise estimates, potentially advancing the frontiers of ultra-high field (UHF) body imaging at 7T and beyond.

Nursing profession during the pandemic: Perceptions, changes, and the role of media

The onset of the COVID-19 pandemic has dramatically highlighted the critical role of nursing in healthcare, transforming the traditional image of nurses from "sisters" or "white angels" to a more esteemed and academic profession. This study aims to evaluate nurses' perspectives on how the media influences the perception of the nursing profession during the COVID-19 crisis and its impact on their professional image. The study seeks to assess nurses' views on the media's role in shaping public perception of nursing during the COVID-19 pandemic and its implications for professional identity. A total of 208 nurses employed in a hospital setting between January 2022 – March 2022 participated in this research, with institutional approval. Data were collected using a researcher-designed questionnaire and the Nursing Profession Image Scale. Statistical analysis was performed using SPSS 22.0, employing descriptive statistics, Pearson correlation, and Student t-tests, with a significance level set at p&lt;0.05. The average age of the nurses was 35.02±7.80 years, with 81.7% being female. Only 7.3% were affiliated with a professional association. A notable 26.0% believed there were positive shifts in the nursing profession's image during the pandemic, while 80.3% observed increased discourse around nursing due to the pandemic. Moreover, 41.8% felt the media adequately portrayed nurses and their professional challenges. The mean score on the Nurses' Image Scale was 112.00±9.67. According to gender, affiliation with a professional organization, and perceptions of media coverage, the perception of the nursing profession's image has varied. However, no significant correlations were found between age, education, professional experience, and perception of the nursing profession. Nurses' perceptions of their professional image are moderately positive. Efforts to enhance the profession's public perception, particularly through media representation, are deemed imperative.

Trends in body image among Australian queer men during COVID-19 lockdowns

ABSTRACT Existing literature demonstrates that COVID-19 lockdowns negatively impacted body image for most people. However, the exact nature of body image shifts during lockdowns and the impact sexuality had on these shifts remains understudied. We sought to investigate this phenomenon among queer men – an at-risk population for negative body image. Participants were 268 (M age = 35.30 years) queer Australian men. They completed measures for lockdown-induced body dissatisfaction, disordered eating and exercising, dysmorphic symptomology, anxiety, depression, and suicidality. Increases in body dissatisfaction and disordered eating during lockdowns were associated with increases in anxiety, depression, and suicidality symptomology. Furthermore, the relationship between body dissatisfaction and depression was moderated by disordered eating and exercise, whereby body dissatisfaction predicted depression only for participants who engaged in heightened disordered eating and overexercise during lockdowns. In other words, body dissatisfaction was most harmful to queer men who specifically sought to remedy this dissatisfaction by engaging in problematic eating/exercise behaviours. These findings have clinical implications for supporting queer men with negative body image in the wake of COVID-19 lockdowns. Namely, queer men might benefit from therapies that build on body neutrality frameworks to help normalise experiences of body dissatisfaction and discourage subsequent disordered eating and exercise.

The Transition of the United States from World Leader to Hegemony During the Cold War

This text scrutinizes how, during the span of the Cold War, particularly at the commencement of hostilities, the United States metamorphosed from a world leader to a dominion of hegemonic might. It perceives the Cold War as a period of ideological corruption within a nation’s ethos. The article first assesses the United States’ geopolitical status before World War II and signs of its role as a world leader, then discusses the role and performance of the United States as a world leader through its actions during and after World War II. Subsequently, it analyzes the transition to U.S. hegemony, using Modelski’s model and Wallerstein’s theory to study the shift in U.S. global strategy and influence, with key events such as the Marshall Plan. The article also examines the ideological shift within the United States during this period, including the impacts of policies from Roosevelt to Truman to Eisenhower. This study provides a comprehensive understanding of key shifts in the geopolitical image of the United States during the Cold War. It bridges geopolitical analysis with critical theory, providing fresh insights into the subtleties of U.S. foreign policy and its global impact.

Abstract PO5-22-02: Evaluation of a surface-guided and tattoo-less approach to setup for whole breast radiotherapy

Abstract Background: In preparation for adjuvant breast radiation therapy (RT), several small but permanent skin tattoo marks are often placed to assist with surface alignment for treatment setup and delivery. With advancements in surface-guided technologies, it may be possible to achieve setup accuracy with a “tattoo-less” approach, which can improve cosmetic and quality of life outcomes. This study evaluates the feasibility, accuracy, and potential dosimetric implications of a tattoo-less setup technique. Method: A retrospective analysis of 10 patients diagnosed with right-sided breast cancer who received adjuvant whole breast RT using a tattoo-less approach were included in this study. All patients received an initial whole breast dose of 42.56 Gray (Gy) in 16 daily fractions using a tangential three-dimensional conformal planning technique. This was followed by a 10Gy boost to the lumpectomy site delivered in 4 daily fractions. Daily setup for whole breast treatment was performed by aligning each patient to their individual reference breast surface generated from the planning computed tomography (CT) images via the Vision RT system. Bony landmarks were then verified via daily kV imaging, and translational shifts were recorded for each fraction. The kV imaging was performed to confirm the treatment positioning since the tattoo-less approach was new to the department. For 5 patients, dosimetric differences between initial surface alignment and shifts performed based on bony landmarks were evaluated by recalculating the original clinical plan with a new isocenter position using kV image shifts for each of their 16 treatment fractions. A plan sum of the 16 recalculated plans for each patient was used in further dosimetric analysis. Boost fractions were not included in this study, as setup was verified using surgical clips instead of bony anatomy. Results: Translational shifts for 158 fractions were reviewed as no kV imaging was performed for the other 2 fractions. The mean absolute kV imaging shift was 0.29 cm (CI (95%): 0.25, 0.32; range: 0 to 1.25) in the vertical axis, 0.28 cm (0.23, 0.32; 0 to 1.27) in the longitudinal axis, and 0.24 cm (0.20, 0.28; 0 to 1.98) in the lateral axis. Clinical plans for 5 patients were evaluated with a mean absolute kV shift of 0.27 cm (range: 0.15 to 0.39) in the vertical, 0.32 cm (0.13 to 0.77) in the longitudinal, and 0.29 cm (0.18 to 0.38) in the lateral directions. Dosimetric analysis revealed a slight increase in mean heart dose from the original clinical plan, with an average difference of 4.12 cGy (range: 0.7 to 10.5 cGy). The maximum increase in V20Gy of the ipsilateral lung was 2.7%. Less than 1% of change was observed in the maximum plan dose, as well as in the V95% coverage, expect for one patient, whose V95% was decreased from 99.6% to 94.6%. This difference was caused by a large longitudinal shift which could be attributed to possible patient movement. Conclusions: These findings demonstrate the feasibility and accuracy of a tattoo-less treatment approach in a small cohort of breast cancer patients receiving whole breast RT. Further investigation with larger cohorts is warranted. Citation Format: Xinxin Zhang, Swati Mamidanna, Yin Zhang, Xiao Wang, Ning Yue, Shicha Kumar, Lindsay Potdevin, Maria Kowzun, Mridula George, Lara Hathout, Bruce Haffty, Nisha Ohri. Evaluation of a surface-guided and tattoo-less approach to setup for whole breast radiotherapy [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO5-22-02.

Exploring French language and culture instruction at an Algerian university: insights from L1 students’ perspectives

Nowadays, it is imperative to comprehend the relationship between intercultural communication and international representations of cultures and nations. This study focuses on exploring the representations that Junior Algerian students hold concerning the French language as well as culture, in order to apprehend students’ attitudes towards the language in question, identify pre-existing stereotypes, and eventually promote a more competent understanding of the culture associated with the language. Through surveys’ responses by 130 juniors at the University of Oum El Bouaghi who have undergone a year of training in French Civilization, researchers attempt to comprehend the impact of the programs followed on the potential reconfiguration of these representations and to identify the relationship of these future educators with the French language. The researchers adopt a descriptive-analytic approach to evaluate the effects of this on students’ conceptions and associated images, concerning the French culture as a whole. The responses exhibited less positive associations with France compared to the French language itself, revealing a positive shift in stereotypical images of French culture among university students. Adopting an intercultural dimension to investigate this underscores the need to prepare learners culturally as well as linguistically in order to effectively interact with different cultures.

Hierarchical alignment network for domain adaptive object detection in aerial images

Domain Adaptive Object Detection (DAOD) alleviates the reliance on labeled data by transferring knowledge learned from labeled source domain to unlabeled target domain. Recent DAOD methods is modeled mainly based on ground-level images. Compared to ground-level images, aerial images suffer from scale variation and viewpoint diversity. This means that domain adaptive object detection in aerial images is a more challenging task. In this work, we address the domain shift in aerial images from two levels: 1) image-level shifts, such as weather, lighting, viewpoint, etc., 2) instance-level shifts, such as object appearance, scale, etc. Specifically, multiple domain-confusion classifiers are designed to learn image-level common knowledge of the source and target domains. Different levels of domain classifiers are further assigned adaptive weights to coordinate the transferability and discriminability of the adaptive detectors. Meanwhile, instance-level alignment is realized by forcing the intrinsic relationship between classes in both domains to be consistent. In addition, we perform instance-level alignment in different semantic-level feature layers to improve the scale awareness of the adaptation model. Extensive experimental results on VisDrone, UAVDT, DIOR and DOTA datasets demonstrate that our method achieves optimal detection performance in four domain adaption scenarios compared to other state-of-the-art methods, e.g., in Daytime → Night (VisDrone), the mAP50 is 23.5 %; in VisDrone → UAVDT, DIOR → UAVDT and DOTA → VisDrone, the AP50 of the car is 63.1 %, 46.6 % and 44.8 % respectively. Code will be available online (https://github.com/MaYou1997/HANet).

Reduction of cervical vertebra ghost images in panoramic radiography using vertical dual exposure.

To evaluate the image quality of vertical dual-exposure panoramic radiography (PR), which merges two PR images taken at different focus heights to reduce ghost images of cervical vertebrae (CV) and intervertebral spaces (IVS) in the incisor region. PR images of an aluminum block, a CV phantom and a human head phantom were taken at 0 mm and merged with and subtracted from PR images taken at other heights (0, 5, 10, 15, and 20 mm) to create new images, e.g., Merg0 + 15 mm and Sub0 - 10 mm. The subtracted images were analyzed subjectively according to the uniformity on the line profile. Merged images were evaluated subjectively by six raters to determine the influence of the ghost images. Objective evaluation revealed a positional shift in the ghost images according to the height of the focus for both phantoms. In the subjective evaluation, the normal PR (Merg0 + 0 mm) showed the worst score, indicating strong influence of CV and IVS ghost images. The vertical dual-exposure PR method, which merges PR images taken at the normal position and a higher X-ray focus, can reduce CV and IVS ghost images in the incisor region.

Stomping on Sacred Grounds

Abstract This paper investigates the changes in the uncanny expressed in historical examples of sumo wrestling. Various pre-modern instances show sumo as a wild and relentless combat, often resulting in the injury or death of one of the wrestlers. However, due to sociopolitical changes in the early Edo period (1600–1868), a new mode of conduct emerged at shrines and temples that complied with regulations in the urban space and forced wrestling to adapt and curtail the uncanny. Eventually, a further drastic image shift occurred when popular culture picked up sumo wrestling. Finally, in the latter half of the nineteenth century, sumo wrestling turned into a “sport” while still retaining uncanny elements in its conduct. Thus, this paper concludes that religious institutions tamed sumo wrestling’s classical uncanniness; at the same time, new conduct secured the integration of uncanny elements within modern sumo wrestling.

Comparing the Robustness of ResNet, Swin-Transformer, and MLP-Mixer under Unique Distribution Shifts in Fundus Images.

Diabetic retinopathy (DR) is the leading cause of visual impairment and blindness. Consequently, numerous deep learning models have been developed for the early detection of DR. Safety-critical applications employed in medical diagnosis must be robust to distribution shifts. Previous studies have focused on model performance under distribution shifts using natural image datasets such as ImageNet, CIFAR-10, and SVHN. However, there is a lack of research specifically investigating the performance using medical image datasets. To address this gap, we investigated trends under distribution shifts using fundus image datasets. We used the EyePACS dataset for DR diagnosis, introduced noise specific to fundus images, and evaluated the performance of ResNet, Swin-Transformer, and MLP-Mixer models under a distribution shift. The discriminative ability was evaluated using the Area Under the Receiver Operating Characteristic curve (ROC-AUC), while the calibration ability was evaluated using the monotonic sweep calibration error (ECE sweep). Swin-Transformer exhibited a higher ROC-AUC than ResNet under all types of noise and displayed a smaller reduction in the ROC-AUC due to noise. ECE sweep did not show a consistent trend across different model architectures. Swin-Transformer consistently demonstrated superior discrimination compared to ResNet. This trend persisted even under unique distribution shifts in the fundus images.

Exit wave function reconstruction from two defocus images using neural network

Wave function reconstruction from one or two defocus images is promising for live atomic resolution imaging in transmission electron microscopy. However, a robust and accurate reconstruction method we still need more attention. Here, we present a neural-network-based wave function reconstruction method, EWR-NN, that enables accurate wave function reconstruction from only two defocus images. Results from both simulated and two different experimental defocus series show that the EWR-NN method has better performance than the widely-used iterative wave function reconstruction (IWFR) method. Influence of image number, defocus deviation, residual image shifts and noise level were considered to validate the performance of EWR-NN under practical conditions. It is seen that these factors will not influence the arrangement of atom columns in the reconstructed phase images, while they can alter the absolute values of all-atom columns and degrade the contrast of the phase images.

Dense speed-of-sound shift imaging for ultrasonic thermometry

Objective. Develop a dense algorithm for calculating the speed-of-sound shift between consecutive acoustic acquisitions as a noninvasive means to evaluating temperature change during thermal ablation. Methods. An algorithm for dense speed-of-sound shift imaging (DSI) was developed to simultaneously incorporate information from the entire field of view using a combination of dense optical flow and inverse problem regularization, thus speeding up the calculation and introducing spatial agreement between pixels natively. Thermal ablation monitoring consisted of two main steps: pixel shift tracking using Farneback optical flow, and mathematical modeling of the relationship between the pixel displacement and temperature change as an inverse problem to find the speed-of-sound shift. A calibration constant translates from speed-of-sound shift to temperature change. The method performance was tested in ex vivo samples and compared to standard thermal strain imaging (TSI) methods. Main results. Thermal ablation at a frequency of 2 MHz was applied to an agarose phantom that created a speed-of-sound shift measured by an L12-5 imaging transducer. A focal spot was reconstructed by solving the inverse problem. Next, a thermocouple measured the temperature rise during thermal ablation of ex vivo chicken breast to calibrate the setup. Temperature changes between 3 °C and 15 °C was measured with high thermometry precision of less than 2 °C error for temperature changes as low as 8 °C. The DSI method outperformed standard TSI in both spatial coherence and runtime in high-intensity focused ultrasound-induced hyperthermia. Significance. Dense ultrasonic speed-of-sound shift imaging can successfully monitor the speed-of-sound shift introduced by thermal ablation. This technique is faster and more robust than current methods, and therefore can be used as a noninvasive, real time and cost-effective thermometry method, with high clinical applicability.